INDEX

Explanations

please followed by a verb

This neuron is specialized to detect the period token when it’s used as a list‐item or section number delimiter (e.g. the “.” after numerals like “1.”, “2.”, “3.”).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

etheless

0.30

 them

0.28

 window

0.28

0.27

 dept

0.27

 করলেও

0.27

ooz

0.27

𝘬

0.27

 backside

0.27

ন্ধে

0.26

POSITIVE LOGITS

Ü

0.37

Công

0.34

Ố

0.34

 Công

0.34

 Sản

0.34

À

0.33

Hvis

0.33

Você

0.33

 viaggio

0.33

Í

0.32

Activations Density 0.141%